Significance of integrated in silico transmural ventricular wedge preparation models of human non-failing and failing hearts for safety evaluation of drug candidates.

INTRODUCTION: To evaluate the usefulness of in silico assay in predicting drug-induced QTc prolongation and ventricular proarrhythmia, we describe in this study 2-dimensional transmural ventricular wedge preparation model (2D model) of non-failing (non-FH) and failing hearts (FH) based on O'Hara-Rudy dynamic model of human ventricular myocytes.

METHODS: Using the prepared 2D model, we simulated ventricular action potential and recorded electrocardiogram for the non-FH and FH. The FH model was constructed based on differences in mRNA, protein, and/or current levels of ion channels between non-diseased heart and failing heart. To simulate the effects of selected drugs, we incorporated changes in ion channel conductance depending on the IC50 value and Hill coefficient at unbound drug blood concentrations.

RESULTS: Dofetilide concentration-dependently induced QTc prolongation at therapeutic concentration in the 2D model of both non-FH and FH. The QTc prolongation in FH was longer than that in non-FH. These findings are consistent with previously reported clinical data. At supratherapeutic concentration 20nM, dofetilide induced Torsade de Pointes-like arrhythmia in the 2D non-FH model. In contrast, the single ventricular myocyte model did not quantitatively reproduce experimental data due to lack of electrotonic interaction. The simulated QTc change induced by six drugs examined in the IQ-CSRC prospective study was almost equivalent to that recorded in drug-treated healthy volunteers.

DISCUSSION: Our 2D model with or without heart failure faithfully reproduced drug-induced QT prolongation and ventricular arrhythmias, suggesting that the in silico approach is a powerful tool for predicting cardiac safety of drug candidates at preclinical stage.